RE: [9fans] don't shoot me

RE: [9fans] don't shoot me

[Stephen Parker]

[-- Attachment #1.1: Type: text/plain, Size: 217 bytes --]

"Modern C++ Design" (Andrei Alexandrescu) takes this extra step and attempts to
implement patterns using templates.  If you thought C++ had already gone too far
then this book should confirm this.

stephen




[-- Attachment #1.2: Type: text/html, Size: 696 bytes --]

[-- Attachment #2: Re: [9fans] don't shoot me --]
[-- Type: message/rfc822, Size: 2844 bytes --]

[-- Attachment #2.1.1: Type: text/plain, Size: 754 bytes --]

> I don't think it has much to do with XML, but "design patterns"
> are a genuine advancement in the field of software engineering,
> applicable only with support from object-oriented languages,

Come on.  ``Genuine advancement in the field of software engineering''
is a bit strong.  As Gamma et al themselves point out in the
introduction, nothing in the book is new.  The only advancement they
made was to package them up in one place and give names everyone could
use as the basis for a common vocabulary.

> in the case of C++ using templates to implement the patterns.

Funny, most of the design patterns in Gamma et al don't use templates.
Only four or five out of 23 do.  Most just use simple inheritence
hierarchies.

	- Dan C.

[-- Attachment #2.1.2: Type: text/html, Size: 1471 bytes --]

Re: [9fans] don't shoot me

[jared jennings]

On Sat, 19 Jul 2003 03:13:31 +0000, Geoff Collyer wrote:

>> [quoted text muted]
> 
> I haven't used XML, but tree structure seems like a simple thing to
> provide, for example with indentation as Plan 9 prototype files and
> Python do:

i was surprised that no one mentioned YAML (www.yaml.org). it is just such
an indentation thing. it seems cool.

-- 
check out sneakemail.com and fastmail.fm. (they didn't pay me to say this)

Re: [9fans] don't shoot me

[Dan Cross]

> I don't think it has much to do with XML, but "design patterns"
> are a genuine advancement in the field of software engineering,
> applicable only with support from object-oriented languages,

Come on.  ``Genuine advancement in the field of software engineering''
is a bit strong.  As Gamma et al themselves point out in the
introduction, nothing in the book is new.  The only advancement they
made was to package them up in one place and give names everyone could
use as the basis for a common vocabulary.

> in the case of C++ using templates to implement the patterns.

Funny, most of the design patterns in Gamma et al don't use templates.
Only four or five out of 23 do.  Most just use simple inheritence
hierarchies.

	- Dan C.

Re: [9fans] don't shoot me

[boyd, rounin]

LaTeX?  nah won't wear it or use it.  what a god-awful idea.

Re: [9fans] don't shoot me

[boyd, rounin]

> I don't think it has much to do with XML, but "design patterns"
> are a genuine advancement in the field of software engineering,
> applicable only with support from object-oriented languages,
> in the case of C++ using templates to implement the patterns.

well i appreciate its goals, but it does seem to have run off the rails.
i see this slavish application of this lunatic philosophy and well, i
can't say i like it.

programming is an art, it's a long way from being a science.

Re: [9fans] don't shoot me

[Anthony Mandic]

"boyd, rounin" wrote:

> don't get me started on LaTeX ...

	Just won't wear it, eh?

-am	© 2003

Re: [9fans] don't shoot me

[Douglas A. Gwyn]

boyd, rounin wrote:
> if you need a pattern, go knit something.
> if you need code, just code it.

I don't think it has much to do with XML, but "design patterns"
are a genuine advancement in the field of software engineering,
applicable only with support from object-oriented languages,
in the case of C++ using templates to implement the patterns.

The idea, which is a recurring theme in software engineering,
is to leverage the programmer's efforts by enabling more
applications for the same amount of work.  Just as OO
technology used properly avoids having to enter similar code
details for multiple objects with similar properties, so
design patterns used properly avoids having to create similar
code scaffolding (higher-level control logic) to implement
similar functionality for different unrelated object classes.

If you haven't read the book by Gamma et al. you should do so.

Of course, you don't need industrial-strength tools if you're
just tinkering, but still you should appreciate their value.

Re: [9fans] don't shoot me

[Anthony Mandic]

ron minnich wrote:

> so much more readable than lisp.

	More like today's Cobol.

-am	© 2003

Re: [9fans] don't shoot me

[boyd, rounin]

> there is a fine line between the annoyance factor of "being strict"
> and knowing that you are presented with well-formed data.

not known to work in finite time, but the code is huge ...

Re: [9fans] don't shoot me

[Bruce Ellis]

this works for me, but sure the parser took some careful treading
(it's nearly 2000 lines but does pass all conformance tests i've seen).

there is a fine line between the annoyance factor of "being strict"
and knowing that you are presented with well-formed data.

<?xml version="1.0" encoding="UTF-8"?>
<!DOCTYPE xwm SYSTEM "xwm.dtd" [
 <!ENTITY wmpath "/dis/wm">
]>
<xwm>
 <param name="Background" value="26,36,136"/>
 <param name="ToolHeight" value="48"/>
 <param name="Icon" value="inferno.bit"/>
 <exec cmd="&wmpath;/plumb.dis"/>
 <menu>
  <item label="About" value="about"/>
  <submenu label="System">
   <item label="Config" value="&wmpath;/xamal.dis xml/xwm.xml"/>
   <item label="Task Manager" value="&wmpath;/task.dis"/>
   <item label="Limbo Debugger" value="&wmpath;/wmdeb.dis"/>
  </submenu>
  <submenu label="Applications">
   <item label="Coffee" value="&wmpath;/coffee.dis"/>
   <item label="Text Editor" value="&wmpath;/wmedit.dis"/>
   <item label="Mash" value="&wmpath;/wmmash.dis"/>
  </submenu>
  <sep/>
  <item label="Top" value="taskmenu"/>
  <item label="Tasks" value="task"/>
  <item label="Editor" value="edit"/>
  <item label="Shell" value="sh"/>
  <sep/>
 </menu>
</xwm>

Re: [9fans] don't shoot me

[boyd, rounin]

> joe armstrong (erlang guy) has proposed ubf:

Unified Bullshit Fuckup?

it's been a long night, m'lud ...

Re: [9fans] don't shoot me

[Taj Khattra]

On Fri, Jul 18, 2003 at 05:12:35PM +0200, pac wrote:
>
> as an absolute amateur (and before diving into) i would like to know your
> opinion on XML
>

see http://www.research.avayalabs.com/user/wadler/language.pdf  :)

joe armstrong (erlang guy) has proposed ubf:

	http://www.sics.se/~joe/ubf/site/home.html

-taj

Re: [9fans] don't shoot me

[boyd, rounin]

looks a bit like bib, no?

Re: [9fans] don't shoot me

[Geoff Collyer]

Labelling is compatible with my suggestion.  One possibility would be
to use NOV extension-field syntax without the colons:

address	street 600 Mountain Ave.	city Murray Hill	state NJ	postcode 07974-2070	country USA

or something that more obviously permits arbitrary substructure:

address
	country USA
	postcode 07974-2070
	state NJ
	city Murray Hill
	street
		number 600
		name Mountain Ave.

or even reflects the geographical hierarchy of the address:

address
	country USA
		postcode 07974-2070
		state NJ
			city Murray Hill
				street Mountain Ave.
					number 600
						building 2
							section C
								room 501
									person G. R. Emlin

None of these possibilities require the visible punctuation of either
XML or Lisp, and I've come to appreciate the lack of visual noise.
Presotto's wonderful pseudo-XML,

<message type=answer><sentence><word><pronoun flavor=personal pos=subject>I</pronoun></word><word>
<verb flavor=active type=contraction>haven't</verb></word><noun-phrase><word><article flavor=indefinite>
an</article></word><word><noun pos=object>opinion</noun></word></noun-phrase><diacritical>.</diaciritical>
</sentence></message>

strikes me as JCL for the 21st century.

Re: [9fans] don't shoot me

[boyd, rounin]

XML strikes me like MIME; by about the 4th MIME doc Coppola's quote
[re: Apocalypse Now] from Hearts Of Darkness is earily appropriate:

    we had too much money and too much equipment and, slowly, but
    surely, we went insane

which perfectly parallels the <strike>war</strike> police action.

Re: [9fans] don't shoot me

[boyd, rounin]

> btw- for Mathematics, give me LaTeX any day over eqn and troff.  :-)

god no.  given me troff or give me death.

Re: [9fans] don't shoot me

[Dan Cross]

> > Scott Schwartz once proposed using LaTeX ...
>
> don't get me started on LaTeX ...

Well, it's not LaTeX itself he proposed using, just the syntax.
btw- for Mathematics, give me LaTeX any day over eqn and troff.  :-)

	- Dan C.

Re: [9fans] don't shoot me

[boyd, rounin]

> Scott Schwartz once proposed using LaTeX ...

don't get me started on LaTeX ...

Re: [9fans] don't shoot me

[Dan Cross]

> > Yes, but with some tree structure and naming too.
>
> I haven't used XML, but tree structure seems like a simple thing to
> provide, for example with indentation as Plan 9 prototype files and
> Python do:
>
> /
> 	sys
> 		man
> 			1
> 				cat
> 		doc
> 			fonts
> 			9.ms
>
> Obviously each line could contain more than a simple file name
> component for general data representation.

I guess the one cool think about XML that you don't get with something
like this is a way to name the data meaningfully and naturally.  With
XML, you can write something like the following:

<?xml version="1.0"?>
<address>
  <street>123 Main Street</street>
  <city>Cherry Hill</city>
  <state>New Jersey</state>
  <postcode>12345</postcode>
  <country>USA</country>
</address>

Now, looking beyond the hideous syntax for a moment, one thing jumps
out: all the data is clearly labelled.  I know that ``123 Main Street''
is a street address, and I can extract street addresses by name,
instead of relying on it being in some conventional place in the record
(ie, ``the first text field in each record is the street address'').
Is that useful?  Sometimes yes, more often no.  But when it's needed,
it's really needed and is indispensible.

I contend the only real advantage of XML over other representations is
that it forces data to be labelled (you can do this with sexp's, but
it's not mandatory).  Of course, 9 times out of 10, the labelling
sucks and tells you nothing.  The corresponding sexp might look something
like the following, btw:

(address
  (street "123 Main Street")
  (city "Cherry Hill")
  (state "New Jersey")
  (postcode "12345")
  (country "USA"))

But the following:

("123 Main Street" ("Cherry Hill" "New Jersey") ("12345" "USA"))

is also valid.  Unfortunately, the latter example doesn't preserve any
metainformation about the data; we as humans can look at this and say,
``oh, that looks like an address; 12345 is probably a zip code.''  But
a computer has no idea, and I have no way to tell it other than by
position.  But what if I decide to add another field between the street
address and City/State tuple?  Say, an apartment number field?  All of
a sudden, my position-based extraction logic fails.  At least with XML,
that isn't a problem (in theory, anyway; like I said, the labelling can
be totally bonheaded and meaningless).

Scott Schwartz once proposed using LaTeX syntax for describing data
in the same way one uses XML.  It was a good idea, and we'd end up
with something like:

\begin{address}
  \street{123 Main Street}
  \city{Cherry Hill}
  \state{New Jersey}
  \postcode{12345}
  \country{USA}
\end{address}

Of course, that made too much sense and thus never caught on.  It
would have been a lot cleaner and more compact than using XML
syntax, though.

Oh well.  Like anything else, XML has its place, but it's been shoe
horned into 80 billion different places it doesn't belong.

	- Dan C.

Re: [9fans] don't shoot me

[Geoff Collyer]

[-- Attachment #1: Type: text/plain, Size: 374 bytes --]

> Yes, but with some tree structure and naming too.

I haven't used XML, but tree structure seems like a simple thing to
provide, for example with indentation as Plan 9 prototype files and
Python do:

/
	sys
		man
			1
				cat
		doc
			fonts
			9.ms

Obviously each line could contain more than a simple file name
component for general data representation.

[-- Attachment #2: Type: message/rfc822, Size: 3415 bytes --]

From: Dave Lukes <davel@luchie.demon.co.uk>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] don't shoot me
Date: Fri, 18 Jul 2003 16:50:17 +0100
Message-ID: <20030718165017.A1419@luchie.demon.co.uk>

<raise-head-above-parapet>
Speaking as someone who's trained/used/sworn-at XML/XSLT/XSL-FO ...

> > just drop me a few words, or just [1] .. [5]  ;-)

[1] .. [5] :-).

What Trickey said, plus.

Seriously, XML has it's uses.

e.g. XSL-FO may one day be a tolerable troff replacement.
(Although I'd hate to be an implementor of an XSL-FO processor:
 the reference documents sent me to sleep in record time).

On Fri, Jul 18, 2003 at 05:23:25PM +0200, Fco. J. Ballesteros wrote:
> IMHO, it's text data reinvented, with some help to automate parsing.

Yes, but with some tree structure and naming too.

One of the "interesting" "features" I've noticed about XML is the
"if the only tool you have is a hammer, everything looks like a nail"
syndrome.

You end up with XSLT (i.e. XML) programs manipulating XML data
driven by XML configuration files and so ad infinitum.
(Look at the Apache stuff, like ant for an example).

(Yes, I know XSLT can _generate_ plain text, just about,
but it can't _read_ it!)

This disease will be familar to the Lispers out there.

</raise-head-above-parapet>

Cheers,
	Dave.

Re: [9fans] don't shoot me

[boyd, rounin]

> "Design After First Testing, and Constant Refactoring After Programming."

tell me about it ...

Re: [9fans] don't shoot me

[Skip Tavakkolian]

> well then there's 'extreme programming' and 'patterns'.

These guys have come up with a phrase that captures the XP methodology
and has an apropos acronym:

http://www.softwarereality.com/ExtremeProgramming.jsp


"Design After First Testing, and Constant Refactoring After Programming."

Re: [9fans] don't shoot me

[boyd, rounin]

> One catch: the numbers in the problems were written with Roman
> numerals.

nice

Re: [9fans] don't shoot me

[boyd, rounin]

> The stench of their bible
> thumping has tainted it and we've ended up with some real monstrosities
> (like namespaces; ew).

well then there's 'extreme programming' and 'patterns'.  the ontly thing
'extreme' about
that nonsense is that i want to terminate those zealots with _extreme_
prejudice.

if you need a pattern, go knit something.

if you need code, just code it.

Re: [9fans] don't shoot me

[Dan Cross]

> "Why not XML?

Why not, indeed.  When Paul Lustgarten was at Avaya, he was giving a
talk on natural language processing.  He had developed his own
metalanguage and support scaffolding to describe conversations, and
knew one of the questions that would come up was, ``why didn't you use
XML'' (replace XML with VoxML, or whatever DTD du jour you want).  His
response was to start the talk off with a, ``why not XML?'' slide, on
which he gave some simple arithmetic problems he asked the audience to
solve.  One catch: the numbers in the problems were written with Roman
numerals.

But back to XML....  I've used XML quite a bit, and also waffled over
it for years.  On one hand, I think the idea of a language-based way to
describe data (and capture its meaning, i.e., this is a street and not
just a text item) is a good one, but on the other, XML implements it
horribly.  First and foremost, it inherits the God awful syntax of SGML
(the true swiss army markup language).  Second, it's been abused to
death.  As various folks have mentioned, it's been applied like deet in
New York City after a rainstorm:  sometimes useful, most often
unnecessary and malodorous.  I think SEXP's are a nice alternative
(even if they are lisp), but unfortunately, no one uses them (I did
write a parser for them once, though).  Then again, no one uses Plan 9
but that doesn't stop me.  Third and finally, like Linux, it has a
small army of rabid prostelytizers following and promoting it who are
convinced it's the cure to world hunger.  The stench of their bible
thumping has tainted it and we've ended up with some real monstrosities
(like namespaces; ew).

	- Dan C.

Re: [9fans] don't shoot me

[ron minnich]

On Fri, 18 Jul 2003, Scott Schwartz wrote:

> | I like XML as an information exchange language for diverse platforms.
> | Being as vast as it is, it no doubt will be abused more often than
> | properly employed.  But I am not aware of any preferable alternative
> | (ASN.1?).
>
> S-expressions.
>

yep. see: http://www.acl.lanl.gov/supermon/ for our monitoring tools that
are all SEXP based, and have a self-describing data format. The data
streams are composable, since they are S-expressions, so combining data
from 1024 streams to one stream is pretty easy. They include a kernel
monitor that produces Linux /proc type information in S-expression format,
with the difference that our module accomodates lots more data and lots
more information than Linux provides normally.

For a small, fast, embeddable library that we wrote here
see:http://sexpr.sourceforge.net/. This has found non-Supermon uses all
over the place, including in a telescope somewhere. There is also a "Why
not XML" section on that page.

In the limit, and I know this sounds nuts, I kind of wish everything in
/proc (and /dev/*/whatever) produced SEXP. While Plan 9 is nowhere near as
inconsistent as Linux in the /proc area, the structure is still pretty
inconsistent. Some things are lists, others are ordered pairs, and others
are tables. If it were all SEXP it could be self-describing and uniformly
parseable with the small, fast library. But maybe I am nuts.

ron
p.s. Betcha this mail gets filtered due to the SEXP content. But that's
what they're called by old-time lisp dudes.

p.p.s. OH well, I'll quote Matt:

"Why not XML?
The goal of any project should be to solve a problem in an elegant, well
designed manner. Buzzword compliance rarely achieves this, other than
allowing a lowly programmer or software engineer to appease higher-ups who
read that "XML is the latest great thing!" XML suffers from certain flaws
that make it worthless in an environment where speed and efficiency
(memory and time complexity) are key requirements. Compare the
performance, footprint, and developer complexity of Xerces and this
s-expression library, and you'll see that for hierarchical data
representation, one is significantly simpler than the other without losing
the ability to do its job (represent data!). After working with XML for a
while (~3-4 years) on multiple projects, I find that since starting to use
s-expressions again as a data format has allowed me to spend less time
debugging my data representation code and more time on problems that
matter. XML sucks, enough said."

Re: [9fans] don't shoot me

[Scott Schwartz]

| I like XML as an information exchange language for diverse platforms.
| Being as vast as it is, it no doubt will be abused more often than
| properly employed.  But I am not aware of any preferable alternative
| (ASN.1?).

S-expressions.

RE: [9fans] don't shoot me

[Nigel Roles]

>
> so much more readable than lisp.
>
> ron

And slightly less readable than binary.

Re: [9fans] don't shoot me

[boyd, rounin]

XML?  swiss army markup language.

Re: [9fans] don't shoot me

[matt]

cat i_love.You.xml >/dev/brain
ERROR: you say you love me but then you hit me
ERROR: does not compute
fzzzttt
Quick Captain, let's make our escape

Re: [9fans] don't shoot me

[ron minnich]

On Fri, 18 Jul 2003, David Presotto wrote:

> <message type=answer><sentence><word><pronoun flavor=personal
> pos=subject>I</pronoun></word><word> <verb flavor=active
> type=contraction>haven't</verb></word><noun-phrase><word><article
> flavor=indefinite> an</article></word><word><noun
> pos=object>opinion</noun></word></noun-phrase><diacritical>.</diaciritical>
> </sentence></message>
>

so much more readable than lisp.

ron

Re: [9fans] don't shoot me

[steve.simon]

I agree entirely with the sentiments of the below :-)

BTW, This has gotta be a fortune.

-Steve

> <message type=answer><sentence><word><pronoun flavor=personal pos=subject>I</pronoun></word><word>
> <verb flavor=active type=contraction>haven't</verb></word><noun-phrase><word><article flavor=indefinite>
> an</article></word><word><noun pos=object>opinion</noun></word></noun-phrase><diacritical>.</diaciritical>
> </sentence></message>

Re: [9fans] don't shoot me

[David Presotto]

<message type=answer><sentence><word><pronoun flavor=personal pos=subject>I</pronoun></word><word>
<verb flavor=active type=contraction>haven't</verb></word><noun-phrase><word><article flavor=indefinite>
an</article></word><word><noun pos=object>opinion</noun></word></noun-phrase><diacritical>.</diaciritical>
</sentence></message>

Re: [9fans] don't shoot me

[Dave Lukes]

<raise-head-above-parapet>
Speaking as someone who's trained/used/sworn-at XML/XSLT/XSL-FO ...

> > just drop me a few words, or just [1] .. [5]  ;-)

[1] .. [5] :-).

What Trickey said, plus.

Seriously, XML has it's uses.

e.g. XSL-FO may one day be a tolerable troff replacement.
(Although I'd hate to be an implementor of an XSL-FO processor:
 the reference documents sent me to sleep in record time).

On Fri, Jul 18, 2003 at 05:23:25PM +0200, Fco. J. Ballesteros wrote:
> IMHO, it's text data reinvented, with some help to automate parsing.

Yes, but with some tree structure and naming too.

One of the "interesting" "features" I've noticed about XML is the
"if the only tool you have is a hammer, everything looks like a nail"
syndrome.

You end up with XSLT (i.e. XML) programs manipulating XML data
driven by XML configuration files and so ad infinitum.
(Look at the Apache stuff, like ant for an example).

(Yes, I know XSLT can _generate_ plain text, just about,
but it can't _read_ it!)

This disease will be familar to the Lispers out there.

</raise-head-above-parapet>

Cheers,
	Dave.

RE: [9fans] don't shoot me

[Trickey, Howard W (Howard)]

Lucio De Re wrote:
> I like XML as an information exchange language for diverse platforms.
> Being as vast as it is, it no doubt will be abused more often than
> properly employed.  But I am not aware of any preferable alternative
> (ASN.1?).

I agree with this.  ASN.1 is an alternative, but it sucks much worse.

But it is disappointing that the XML spec is so large that you can't just whip up a parser for the whole thing in a few hundred lines. I think there are three reasons for this:

1) it tries to represent two quite different kinds of data: marked up text (like HTML, SGML), and tree-structured data.  The need for the former introduces all sorts of whitespace worries that just muck things up for the latter.
2) 'attributes' are perhaps a convenient shorthand, but could have (and should have, in my opinion) been left out
3) the namespace stuff vastly complicates things for not a lot of gain

- Howard

Re: [9fans] don't shoot me

[Fco.J.Ballesteros]

IMHO, it's text data reinvented, with some help to automate parsing.

Re: [9fans] don't shoot me

[ron minnich]

On Fri, 18 Jul 2003, pac wrote:

> as an absolute amateur (and before diving into) i would like to know
> your opinion on XML (the voices vary from the "[1]best thing since the
> invention of pre-sliced bread" and "eXtremely Messed-up Language[5]"

microsoft really likes XML.

ron

Re: [9fans] don't shoot me

[Lucio De Re]

On Fri, Jul 18, 2003 at 05:12:35PM +0200, pac wrote:
>
> as an absolute amateur (and before diving into) i would like to know your
> opinion on XML
> (the voices vary from the "[1]best thing since the invention of pre-sliced
> bread" and "eXtremely Messed-up Language[5]"
>
I like XML as an information exchange language for diverse platforms.
Being as vast as it is, it no doubt will be abused more often than
properly employed.  But I am not aware of any preferable alternative
(ASN.1?).

++L

[9fans] don't shoot me

[pac]

friends,

as an absolute amateur (and before diving into) i would like to know your
opinion on XML
(the voices vary from the "[1]best thing since the invention of pre-sliced
bread" and "eXtremely Messed-up Language[5]"

just drop me a few words, or just [1] .. [5]  ;-)

sorrry for just another survey,
(yes, and i like the idea, but not the "verbose" implementation)
cheers,

++pac.